New project for developing self-learning XAI-based maintenance system in industry
The Swedish manufacturing industry faces a number of challenges in an increasingly digitised and competitive world. To expand productivity and growth, technology such as AI (artificial intelligence) and ML (machine learning) is becoming increasingly common. MDU has recently been granted funding for a new project which aims to develop a self-monitoring, self-learning and self-explanatory maintenance system to predict needs and deviations in manufacturing and production processes at industrial companies.
“The goal of the project is to develop a digital twin – a copy in a computer environment of something that exists in real life, such as a machine or production process – for cognitive predictive maintenance. With the help of XAI (Explainable AI) and ML, it will help improve maintenance in manufacturing and production processes,” says Mobyen Uddin Ahmed, member of the project at MDU.
Predicts maintenance and reduces operational downtime
Analysing data from sensors on machines and equipment to predict maintenance needs, known as predictive maintenance, has increased through AI and ML in the manufacturing industry. Predictive maintenance not only predicts what maintenance is required in production processes, it also helps to avoid unnecessary costs such as time, energy and resources. It can also predict deviations and breakdowns in machines. Thus, it can create better conditions for Swedish industry to operate more sustainably.
Collaboration to strengthen Swedish industry
The project, called CPMXai (Cognitive Predictive Maintenance and Quality Assurance using EXplainable AI and Machine Learning), takes place in collaboration with leading operators such as Hitachi High-Tech Europe, SPM Instrument AB, Nordic Electronic Partner, GKN Driveline Köping AB, Adopticum, RISE Research Institutes of Sweden and MDU (as Coordinator), where everyone contributes with various experiences and skills in the area. This cooperation will help strengthen the existing partnership between industry, universities, research institutes and innovators in Sweden.
The results of the project will provide a basis for a platform which enables Swedish industrial companies to start working with Industry 4.0, where predictive maintenance, AI and ML are obvious parts.
“Within the framework of this project, we will be able to develop a scalable solution to meet the specific needs of different companies in terms of predictive maintenance. This in turn enables the Swedish manufacturing industry to become more competitive,” says Shahina Begum, Project manager at MDU.
The project is financed by Vinnova’s Production2030 programme.